I believe that use of AEC is a per app policy decision and a control that should be exposed to apps in an API, not as a device specific [hidden] feature without app control/choice or even knowledge by the app that this processing is performed. Some effects are perfectly appropriate as globally applied for each logical audio device and should not be under application control, like for instance effects that compensate for hardware design decisions such as Microphone compensation algorithms, Room Correction algorithms, etc. Some effects could provide a good user experience implemented either in the app space or in the IHV controlled areas of the Windows audio streaming path, such as speaker virtualization or non-PCM encoding. I suppose there are different philosophies at play here. If you want to try to do AEC in your driver or filter driver you still can. Windows is currently moving in the direction of providing more audio path discoverability, control and flexibility to users and applications. Certain portions of the historically hidden (from apps in kernel mode drivers or filter drivers) IHV value-add globally affecting the experience each Windows application is trying to achieve/control just fit better in the application graph in that vision. Also, I am not sure who the client is in your scenario. Our clients are the Windows users using applications running on Windows and if I as a user use an application that enables me to control the use of the audio processing algorithms that the application makes use of to create the experience for me then me as a client would be perfectly content. It is a paradigm shift for sure but one for the better IMHO. Sincerely, Hakon Strande PM Integrated, Internal, External, and Wireless Audio Devices MediaTech/DMD/Windows Client/Microsoft ________________________________ From: wdmaudiodev-bounce@xxxxxxxxxxxxx [mailto:wdmaudiodev-bounce@xxxxxxxxxxxxx] On Behalf Of Tim Roberts Sent: Thursday, September 28, 2006 4:40 PM To: wdmaudiodev@xxxxxxxxxxxxx Subject: [wdmaudiodev] Re: Vista Audio Architecture Hakon Strande wrote: All legacy driver models should continue to work for basic streaming in most cases but AEC and Mic Array processing are application graph DMO effects in Vista (and hosted by the RTC API as well) and does not belong in the driver layer in our vision of Windows and our desire to enable more powerful and flexible Windows application experiences going forward. I understand your goal here, but how does this philosophy reconcile with a client's quite reasonable desire to have his echo cancellation operate globally over all audio producers and consumers? It's not clear to me that an application-centric focus, using DMO effects in an app graph, can provide the same end-to-end control that one has in XP. -- Tim Roberts, timr@xxxxxxxxx Providenza & Boekelheide, Inc.